Intense 2:7 μm and broadband 2:0 μm emissionfrom diode-pumped

Er3�=Tm3�=Ho3�-doped fluorophosphate glassYing Tian,1,2 Rongrong Xu,1,2 Lili Hu,1 and Junjie Zhang1,*

1Key Laboratory of Materials for High Power Laser, Shanghai Institute of Optics and Fine Mechanics,Chinese Academy of Sciences (CAS), Shanghai 201800, China

2Graduate School of the Chinese Academy of Sciences (CAS), Beijing 100039, China*Corresponding author: jjzhang@mail.siom.ac.cn

Received March 24, 2011; revised July 12, 2011; accepted July 28, 2011;posted July 28, 2011 (Doc. ID 144764); published August 12, 2011

This Letter reports intense emission at 2:7 μm and broadband emission at 2:0 μm from Er3þ=Tm3þ=Ho3þ-doped fluor-ophosphate glass. The fluorescence characteristics and energy transfer upon excitation of a conventional 980nmlaser diode are investigated. Based on the fluorescence spectra and lifetime measurement, the effect of Tm3þand Ho3þ ions on intense 2:7 μm emission in fluorophosphate glass is demonstrated. It is also found that the effectivebandwidth of 2:0 μm emission due to Tm3þ and Ho3þ ions can reach as high as 196nm. These results indicate thatthe advantageous spectroscopic characteristics of Er3þ=Tm3þ=Ho3þ triply doped fluorophosphate glass together withthe outstanding thermal properties may become an attractive host for the mid-IR solid state lasers. © 2011 OpticalSociety of AmericaOCIS codes: 160.5690, 300.6280, 160.4670, 160.2750, 160.3130.

Recently, the development of mid-IR diode-pumped solidstate lasers has been an area of great interest due to theirvarious useful applications including military, remotesensing, eye-safe laser radar, and medical surgery [1,2].Er3þ-doped glass is well suited for the above applicationsowing to its 2:7 μm emission, which is close to the mostpronounced absorption band of water at 3 μm. However,the laser characteristics of 2:7 μm due to the 4I11=2 →4I13=2 transition are rather poor. The intrinsic difficultyto obtain laser action on this transition is mainly relatedto shorter fluorescence lifetime of the upper laser level4I11=2 than that of the lower laser level 4I13=2 [3,4]. To ob-tain intense 2:7 μm emission, the artificial depletion of the4I13=2 level is required, which can be achieved by suitablecodoping of rare-earth ions and utilizing the energy trans-fer between levels of the respective ions having equal orclose energies [5]. It has been reported that Tm3þ ionscan deplete the Er3þ:4I13=2 level to obtain 2:7 μm emissionin GeGaAsS glass using a titanium–sapphire tunable laseras an excitation source, which can be attributed to theenergy transfer between the Tm3þ:3F4 and Er3þ:4I13=2levels [4]. As is shown in Fig. 1, the energy gap betweenthe Er3þ:4I13=2 and Ho3þ:5I7 levels is as small as that be-tween the Er3þ:4I13=2 and Tm3þ:3F4 levels. It is expectedthat depletion by both Ho3þ and Tm3þ ions would pro-vide a more efficient energy transfer route and intense2:7 μm emission than deactivation by Tm3þ alone. Butfew results have been reported on 2:7 μm emission inEr3þ=Tm3þ=Ho3þ-doped glasses pumped by a common980 nm laser diode (LD).Additionally, Ho3þ lasers are also well suited for mid-

IR range-finding, atmospheric monitoring, and sensingapplications because their 2:0 μm emissions are withinan atmospheric transparency window [6,7]. To enhancepump absorption, Ho3þ-doped glasses have been sensi-tized with either Tm3þ, Yb3þ, or Er3þ [8,9]. Nevertheless,there are few reports about the 2:0 μm emission inEr3þ=Tm3þ=Ho3þ-doped glass [10].

In order to get powerful mid-IR emissions fromEr3þ=Tm3þ=Ho3þ-doped glass, the host glass is as impor-tant as the choice of rare-earth ions. Regarding the emis-sion at 2:7 μm, most research has focused on Er3þ-dopedfluoride and chalcogenide glasses [11–13]. Fluoropho-sphate (FP) glasses combine significant advantages ofthe fluoride and phosphate glass, and own a less complexfabrication route than that of fluoride and chalcogenideglasses [14]. They are characterized by good moistureresistance, high solubility for rare-earth ions, and broadabsorption and emission bands [15,16]. In particular, FPglasses appear interesting for high power solid state lasersdue to their low nonlinear refractive index, which willminimize associated phenomena such as self-focusingand self-phase modulation in high power lasers and en-sure the highest possible laser intensity on targets [15].We have previously demonstrated 2:7 μm emission in cur-rent FP glasses [17]. Additionally, based on our preparedFP glasses doped with Tm3þ and Ho3þ, the 2 μm spectro-scopic properties have already been investigated andshow potential applications for mid-IR laser media [6,14].

Fig. 1. Energy transfer sketch of Er3þ=Tm3þ=Ho3þ-doped FPglass when pumped at 980nm.

3218 OPTICS LETTERS / Vol. 36, No. 16 / August 15, 2011

0146-9592/11/163218-03$15.00/0 © 2011 Optical Society of America

In this Letter, we demonstrate intense 2:7 μm andbroadband 2:0 μm emission in Er3þ=Tm3þ=Ho3þ-dopedFP glass. The thermal stability, fluorescence characteris-tics, and relative energy transfer mechanisms are in-vestigated. The development of such material is expectedto generate a powerful laser at 2.7 and 2:0 μm. The in-vestigated glass has the following molar compositions:20AlðPO3Þ3− 80RF2 − 1ErF3 − 1TmF3 − 1HoF3 (R ¼ Mg,Ca, Sr, Ba). The fabrication method as well as thermaland optical property measurements of glass sampleswere described in our previous work [6,14,17]. The sam-ples were fabricated and polished to the size of 20mm ×10mm × 1mm for the optical property measurements.From the thermal analysis, Tg and Tx of the prepared

glass is 416.8 and 638:6 °C, respectively. TheΔTðTx − TgÞhas been frequently used as a rough estimate of the glassstability against crystallization [18]. The glass stability isclosely related to fiber drawing because fiber drawing isa reheating process during which any crystallization willincrease the scattering loss of the fiber and degrade thelaser properties [19]. It is desirable for ΔT to be as largeas possible to achieve a wide working range of tempera-ture during fiber drawing [20]. It is noted that ΔT(221:8 °C) of our present FP glass is significantly higherthan a series of FP glasses (29–90 °C) [21], lead–tellurite–germanate glass (150 °C) [22], and fluoride glass (80 °C)[23]. Thus, the Er3þ=Tm3þ=Ho3þ-doped FP glass is ex-pected to have a better thermal stability during opticalfiber drawing.Figure 2(a) illustrates the absorption spectra of Ho3þ,

Tm3þ, and Er3þ singly and the Er3þ=Tm3þ=Ho3þ-dopedsample in the wavelength region of 390–2100 nm. The ab-sorption bands of Er3þ, Tm3þ, and Ho3þ corresponding totransitions starting from the ground state to higher levelsare labeled. Generally, for the Er3þ=Tm3þ=Ho3þ-dopedsample, the shape and peak positions of each transitionare very similar to those of Er3þ, Tm3þ, and Ho3þ singlydoped samples and there is no shift in the wavelengthof the absorption peaks. This phenomenon indicatesno cluster in the local ligand field and three kinds of ionsare homogeneously incorporated into the glassy net-work. Because of the absorption band around 980 nm,this triply doped sample can be excited efficiently by a980 nm LD. Figure 2(b) presents the IR transmissionspectrum of a triply doped sample with 1mm thickness.

The maximum transmittance reaches as high as 93% andno evident drop of transmittance at 3 μm is observed.Since the typical H2O absorption band at 3 μm would se-verely inhibit laser emission at 2:7 μm [24], the glass formid-IR lasers is expected to possess a high transmittanceat 3 μm. The excellent IR transmission property providesthe Er3þ=Tm3þ=Ho3þ-doped FP glass with potential ap-plications for 2:7 μm laser material.

To minimize the effect of radiation reabsorption,the excitation radiation was launched onto the edge ofthe 1:0mm thick sample and the fluorescence wascollected perpendicularly from the incident beam.Figure 3(a) shows the fluorescence spectrum of theEr3þ=Tm3þ=Ho3þ-doped FP glass in the range of 1700–2150 nm under 980 nm excitation. The 1:8 μm emissionof Tm3þ overlaps with the 2:0 μm emission of Ho3þ,which covers the wavelength in a wide range of 1700–2150 nm. However, the 2.0 and 1:8 μm emissions differgreatly in intensity, as shown in Fig. 3(a). Since the2:0 μm emission band of Tm3þ and Ho3þ ions in glassesis asymmetric, choosing the effective emission band-width Δλeff rather than the FWHM is more reasonable.According to the simplified procedure provided inRef. [25], theΔλeff of 2:0 μm emission in the present glasscan reach as high as 196 nm. Figure 3(b) shows the fluor-escence around 2:7 μm in Er3þ and Er3þ=Tm3þ=Ho3þ-doped FP samples. It was found that in our case, the2:7 μm luminescence spectra at higher pump intensitiesare different in comparison to spontaneous emissionspectra at low excitation intensities. The spectra shownin Fig. 3 were measured at low excitation power (0:18W).Comparing the intensity of the two samples, the intensityof the 2:7 μm emission is weaker in Er3þ singly doped FPglass. It can be attributed to large nonradiative relaxationrates (≈106 s−1) between the 4I11=2 and 4I13=2 levels [26].Therefore, the lower laser level is largely populated,which is detrimental to get 2:7 μm emission [27]. It hasbeen demonstrated that the effective depopulation of the4I13=2 state is necessary for 2:7 μm cw lasers [28], whichcan be achieved by codoping of Tm3þ or Pr3þ to depletethe lower level [4,29]. In Fig. 3(b), 2:7 μm emission ismore intense in the triply doped sample than that of thesingly doped one, which indicates that doping Tm3þ andHo3þ into Er3þ-doped FP glass can be an alternative way

Fig. 2. (a) Absorption spectra of prepared samples. (b) IRtransmission spectrum of Er3þ=Tm3þ=Ho3þ-doped FP glass.

Fig. 3. Diode-pumped (980 nm) fluorescence spectra for(a) broadband 2:0 μm emission of Er3þ=Tm3þ=Ho3þ-doped FPglass and (b) 2:7 μm emission of Er3þ=Tm3þ=Ho3þ and Er3þ sin-gly doped FP glass.

August 15, 2011 / Vol. 36, No. 16 / OPTICS LETTERS 3219

to get 2:7 μm emission. And the peak of the 2:7 μmemission cross section in the triply doped sample wascalculated to be 6:02 × 10−21 cm2, which is comparable toEr3þ=Pr3þ-doped FP glass (6:57 × 10−21 cm2) [17] andEr3þ-doped ZBLAN glass (5:7 × 10−21 cm2) [30].The involved energy transfer mechanisms are indi-

cated in Fig. 1. First, the 4I15=2 level of Er3þ is excited tothe 4I11=2 level by ground state absorption when the sam-ple is pumped by a 980 nm LD. Ions in the 4I11=2 level de-cay radiatively to the 4I13=2 level with 2:7 μm emission ornonradiatively to the 4I13=2 level. It is possible that theEr3þ:4I11=2 level can be deexcited by the Ho3þ:5I6 (ET1)level or the Tm3þ:3H5 (ET2) level. However, when the en-ergy gap (1700 cm−1) of the ET1 process is larger thanthat (1450 cm−1) between the Ho3þ:5I7 and Er3þ:4I13=2 le-vels, quenching is expected to be lower than for theEr3þ:4I13=2 level, as in the similar case reported by Allainet al. [3]. And the energy transfer rate of ET2 has beendemonstrated to be much smaller than that of ET4(Er3þ:4I13=2 → Tm3þ:3F4) [31]. Thus, Tm3þ and Ho3þ canrapidly deplete the lower laser level Er3þ:4I13=2 whilehaving a much lesser effect on the upper laser levelEr3þ:4I11=2, which is beneficial to 2:7 μm emission. Theefficiency of Tm3þ and Ho3þ depopulation can be esti-mated by the lifetime of the Er3þ:4I13=2 level with(1:15ms) and without (8:6ms) Tm3þ=Ho3þ codoping.So the energy transfer efficiency reaches as high as87.08%, which demonstrates that the Ho3þ:5I7 andTm3þ:3F4 levels can efficiently quench the Er3þ:4I13=2 le-vel in the present glass. On one hand, it is beneficial to2:7 μm emission. On the other hand, through the ET3and ET4 processes, the Ho3þ:5I7 and Tm3þ:3F4 levelsare populated. Then 2.0 and 1:8 μm emissions are ob-tained simultaneously due to the Ho3þ:5I7 → 5I8 andTm3þ:3F4 →

3H6 transitions, respectively. However, the2.0 and 1:8 μm emissions differ greatly in intensity, asshown in Fig. 3(a), which results from the energy transferbetween Tm3þ and Ho3þ ions. The energy of Tm3þ:3F4

can be transferred to the Ho3þ:5I7 level (ET5) and, con-sequently, 1:8 μm emission is weakened. As to the upperlevel of 2:0 μm emission (Ho3þ:5I7), it is greatly populatedthrough the ET5 and ET3 processes. As a result, 2:0 μmemission can be enhanced and more intense than that of1:8 μm emission.In conclusion, the Er3þ=Tm3þ=Ho3þ-doped FP glass

with 2:7 μm and broadband 2:0 μm emissions is demon-strated. The thermal stability and spectroscopic proper-ties are characterized. The absence of OH− absorption at3 μm guarantees the observation of intense 2:7 μm emis-sion. When pumped by common a 980 nm LD, codopingTm3þ and Ho3þ with Er3þ in FP glass is an effectual tech-nique for the observation of 2:7 μm emission. Meanwhile,the effective bandwidth of 2:0 μm emissions can reach196 nm. The Er3þ=Tm3þ=Ho3þ-doped FP glass with desir-able thermal resistance properties and spectroscopiccharacteristics is a promising mid-IR laser material.

The authors are thankful to the National NaturalScience Foundation of China (NSFC) (No. 60937003

and No. 50902137) and the GF Innovation Project(No. CXJJ-11-M23 and No. CXJJ-11-S110).

References

1. M. Pollnau and S. D. Jackson, in Solid-State Mid-InfraredLaser Sources, I. T. Sorokina and K. L. Vodopyanov, eds.(Springer, 2003), pp. 219–253.

2. A. S. S. de Camargo, E. R. Botero, E. R. M. Andreeta,D. Garcia, J. A. Eiras, and L. A. O. Nunes, Appl. Phys. Lett.86, 241112 (2005).

3. J. Y. Allain, M. Monerie, and H. Poignant, Electron. Lett. 27,445 (1991).

4. Y. G. Choi, K. Hon Kim, B. Joo Lee, Y. Beom Shin, Y. SongKim, and J. Heo, J. Non-Cryst. Solids 278, 137 (2000).

5. L. Wetenkamp, G. F. West, and H. Többen, J. Non-Cryst.Solids 140, 25 (1992).

6. Y. Tian, L. Zhang, R. Xu, L. Hu, and J. Zhang, Appl. Phys. B101, 861 (2010).

7. Y. Tsang, B. Richards, D. Binks, J. Lousteau, and A. Jha,Opt. Lett. 33, 1282 (2008).

8. Y. Tsang, B. Richards, D. Binks, J. Lousteau, and A. Jha,Opt. Express 16, 10690 (2008).

9. L. X. Yi, M. Wang, S. Y. Feng, Y. K. Chen, G. N. Wang,L. L. Hu, and J. J. Zhang, Opt. Mater. 31, 1586 (2009).

10. M. Wang, L. Yi, G. Wang, L. Hu, and J. Zhang, Solid StateCommun. 149, 1216 (2009).

11. X. Zhu and R. Jain, Opt. Lett. 32, 26 (2007).12. F. Prudenzano, L. Mescia, L. Allegretti, V. Moizan, V.

Nazabal, and F. Smektala, Opt. Mater. 33, 241 (2010).13. A. B. Seddon, Z. Tang, D. Furniss, S. Sujecki, and

T. M. Benson, Opt. Express 18, 26704 (2010).14. Y. Tian, R. Xu, L. Zhang, L. Hu, and J. Zhang, J. Appl. Phys.

109, 083535 (2011).15. D. Ehrt and T. Töpfer, Proc. SPIE 4102, 95 (2000).16. T. Töpfer, J. Hein, J. Philipps, D. Ehrt, and R. Sauerbrey,

Appl. Phys. B 71, 203 (2000).17. Y. Tian, R. Xu, L. Zhang, L. Hu, and J. Zhang, Opt. Lett. 36,

109 (2011).18. S. Xu, Z. Yang, S. Dai, J. Yang, L. Hu, and Z. Jiang, J. Alloys

Compd. 361, 313 (2003).19. L. Le Neindre, S. Jiang, B.-C. Hwang, T. Luo, J. Watson, and

N. Peyghambarian, J. Non-Cryst. Solids 255, 97 (1999).20. M. G. Drexhage, O. H. El-Bayoumi, C. T. Moynihan,

A. J. Bruce, K. H. Chung, D. L. Gavin, and T. J. Loretz,J. Am. Ceram. Soc. 65, c168 (1982).

21. N. Rigout, J. L. Adam, and J. Lucas, J. Non-Cryst. Solids184, 319 (1995).

22. Z. Pan and S. H. Morgan, J. Non-Cryst. Solids 210, 130(1997).

23. M. D. Shinn, W. A. Sibley, M. G. Drexhage, and R. N. Brown,Phys. Rev. B 27, 6635 (1983).

24. S. A. Pollack and M. Robinson, Electron. Lett. 24, 320(1988).

25. M. J. Weber, J. D. Myers, and D. H. Blackburn, J. Appl. Phys.52, 2944 (1981).

26. C. B. Layne, W. H. Lowdermilk, and M. J. Weber, Phys. Rev.B 16, 10 (1977).

27. R. S. Quimby and W. J. Miniscalco, Appl. Opt. 28, 14 (1989).28. T. Sandrock, A. Diening, and G. Huber, Opt. Lett. 24,

382 (1999).29. P. S. Golding, S. D. Jackson, T. A. King, and M. Pollnau,

Phys. Rev. B 62, 856 (2000).30. S. D. Jackson, T. A. King, and M. Pollnau, J. Mod. Opt. 47,

1987 (2000).31. C. Y. Chen, R. R. Petrin, D. C. Yeh, W. A. Sibley, and J. L.

Adam, Opt. Lett. 14, 432 (1989).

3220 OPTICS LETTERS / Vol. 36, No. 16 / August 15, 2011